Five categories of bond specimens have been categorized from the literature: single bar pullout specimens, beam anchorage specimens, beam-end specimens, lap splice tensile, and beam specimens and hook anchorage specimens. Although several experimental studies may be said to use the same type of specimen, the particular details of specimens used in different studies vary.
3.2.1 Pullout specimens
In pullout specimens, generally, the bars are pulled from the surrounding concrete in such a way that the concrete surrounding the bar is subjected to compression (Figure 3.1a). This does not reflect the critical anchorage condition of a bar anchored in a tension zone. In such a test arrangement, compression struts form between the bar surface and the loading points on the concrete surface. Such transverse compression struts cause the effect of increasing the apparent bond strength and are not simulating typical situations encountered in structures and bridges.
Typical in situ conditions have both the bar and the surrounding concrete placed in tension. This type of bond specimen is not recommended by ACI Committee 408 to determine the development length since it represents the least realistic conditions (ACI 2003). A variation of the pullout test often used to test hook anchorages (e.g.: El Hacha et al. 2006) is shown in Figure 3.1b; this specimen suffers from many of the same drawbacks of the single bar pullout specimen.
3.2.2 Beam end specimens
Beam-end specimens represent a more realistic type of specimen that can give better and more accurate results for bond behavior. In the beam end specimen, both the reinforcing and the surrounding concrete are subjected to tensile stresses (Figure 3.1c). To achieve such a state of stress, the compressive forces must be located away from the reinforcing bar a distance not less than the bonded length of the tested bar. Also a short length of the tested bar near the concrete free surface has to be unbonded from surrounding concrete to avoid a conical pullout failure of the concrete. ASTM Standard A944-99 adopts this type of bond specimen. Experimental studies indicate that the bond obtained using beam end specimens closely match those obtained using
full-scale reinforced concrete members (Elagroudy 2003). Ahlborn and DenHartigh (2003) presented a large study on establishing bond properties using beam-end specimens. No beam-end tests are known using high strength reinforcing steel.
3.2.3 Beam anchorage and splice specimens
The beam anchorage and splice specimens represent full scale specimens, designed to directly measure the bond strength of developed or spliced bars. These specimens replicate a flexural member with a defined bond length. Splice specimens are normally designed and tested under four point loading having the splice length lying within the constant moment region (Figure 3.1f). Their ease of fabrication and the close similarity between the stress profile, in both the concrete and reinforcing steel, with the actual stress profile in real flexural members result in splice specimens being the primary source of experimental data used to establish the current design provisions for development length.
Previous studies show that compared with Grade 60 steel, Grade 100 steel allows beam-splice specimens to reach higher loads and deflections before failure (Ansley 2002). Tests also indicate that for bars not confined by transverse reinforcement, longer splices will increase the load at failure and may provide additional ductility, although beyond a certain point increasing the splice length will not increase the load or deformation capacity (Peterfreund 2003).
Shebly (2008) tested full scale beam-splice test specimens in four-point bending having A1035 reinforcement. The results indicate that the development length equation prescribed in ACI 318-05 is not suitable for the design of unconfined or confined splices of this higher grade steel without the use of an additional modification factor. The development length equation recommended in ACI 408R-03 was, however, found to be suitable for both unconfined and
confined splices. Based on this study, both ACI 408 and ACI 318-05 code equations for bond underestimate the bond force capacity at low stress levels in the bar (within the linear portion of the stress-strain curve) and progressively overestimate the bond capacity of A1035 bars when the tensile stress levels exceed the proportional limit. This observation should not be surprising as existing bond recommendations are largely empirical and have been calibrated for steel exhibiting a linear behavior and a having a yield stress less than 75 ksi.
Elagroudy et al. (2006) also tested beam-splice specimens using high-strength steel bars.
They concluded that there is no reason to believe that the bond behavior of the A1035 reinforcing bars is different from that of conventional carbon steel for stress levels below the proportional limit of the A1035 steel (approximately 70 ksi). The nonlinear behavior of the A1035 bars at high stress levels is considered to be the reason behind the observed change in the mode of failure from sudden to gradual. Elagroudy et al. indicate that the nonlinear ductile response of A1035 bars at stress levels beyond the proportional limit, has a strong influence in reducing the bond strength of A1035 bars compared to the bond strength that can be obtained when using other types of steel bars with the same splice length and level of confinement, but with linear stress-strain behavior at high stress levels.
Harries et al. (2010) present research on bond characterization of high-strength steel bars in beam splice specimens. This study clearly demonstrated that the present AASHTO (2007) and ACI 318 (2008) requirements for straight bar tension development length may be extended to develop bar stresses of at least 125 ksi (860 MPa) for concrete strengths up to 10 ksi (69 MPa).
For higher strength steel, greater bar strain and slip will occur prior to development of the bar.
The associated displacement of the bar ribs drives a longitudinal splitting failure beyond that where yielding of conventional bars would occur; thus, confining reinforcement is critical in
developing higher strength bars. The results presented by Harries et al. (2010) and previous work of Seliem et al. (2009) clearly indicate that confining reinforcement should always be used when developing, splicing or anchoring ASTM A1035 reinforcing steel.
3.2.4 Hook anchorage specimens
Figures 3.1d and e represent full scale hook anchorage specimens designed to directly measure the pull-out capacity of hooked bar anchorages. A number of variations of these specimens exist but most mimic the embedment of a beam or girder longitudinal reinforcing bar into a column joint region and are modeled on a study conducted by Marques and Jirsa (1975). This setup allows the developed bar to be located in a concrete tension field and the concrete compressive strut developed by the hook tail to be anchored by an appropriately located reaction mimicking the compressive zone of the beam whose steel is being developed. A variation of this test set-up is adopted in the present study and described in Section 3.4.
Ciancone (2008) evaluated the behavior of standard hook anchorage specimens made using #5 and #7 A1035 steel. No confinement reinforcement was provided in the specimens.
While the #5 hooks were able to develop their yield capacity of 100 ksi, the #7 hooks were not.
This result suggests an effect of bar size and reinforces the need for confining reinforcement when developing A1035 bars.